@kavya
Nothing yet — your first passage lights this up.
Teach Mozhi a little of your Tamil
Original passages — poems, prose, a paragraph in your own voice. Raw, clean Tamil is the fuel; your writing is what grows the model. Each piece is released under CC0.
Saved to your local profile first, then sent to the open corpus when it's reachable. See the contributor terms.
Mozhi is listening.
உங்கள் முதல் வரிகளுக்காகக் காத்திருக்கிறது
Teach it your first lines above — even one honest paragraph of your own Tamil is fuel.
↑ Write your first passageStatuses sync once your contributions reach the corpus DB. Until then everything reads “queued” — honest: nothing is trained yet.
Soon: correct what Mozhi writes
Next, you'll read Mozhi's Tamil and correct it — closing the loop. Mozhi is tiny today (3.2M params, 2.07 bits/char); your writing now is what earns it the right to be corrected later. Sequencing is the point.